Agricultural product candidates are normally identified either by screening test compounds for biological activity on intact plants grown in soil ("greenhouse" assays) or by screening test compounds for inhibition of specific target enzymes or other proteins in vitro ("in vitro" assays). In greenhouse assays, test compounds are either sprayed onto soil containing ungerminated seeds (pre-emergent application) or onto the plants themselves (post-emergent application). Greenhouse assays have a high success rate in predicting the herbicidal activity of a compound when applied in the field. Additionally, a greenhouse assay simultaneously tests for compounds which affect essentially all potential modes of herbicidal action. In other words, greenhouse assays possess a high information content. However, greenhouse testing is time- and space-consuming, requiring 2 weeks and 2 ft.sup.2 of greenhouse space for each compound tested. Large amounts of labor are also required to prepare, care for, spray and score these tests. Moreover, 10 to 60 mg of a test compound is required to assess its potential herbicidal activity at normal use rates. The requirement for this amount of test compound places significant limitations on the synthesis or acquisition of compounds to be tested.
In vitro assays, by contrast, typically assay the affect of a test compound on only a specific target enzyme or protein. These targets are usually extracted proteins, and their responses to the test compounds are assayed in vitro. Alternatively, these targets are expressed in surrogate microorganisms and their responses are assayed by their effects on the growth or metabolism of the microorganisms. In either case, these assays have great advantages over greenhouse assays, including much lower space, labor, time and compound requirements. However, the value of these assays for predicting herbicidal or other biological activity under natural conditions is limited by complex and poorly understood processes unique to intact plants which are absent in vitro or in surrogate organisms. Processes such as uptake, translocation, and metabolism of test compounds cannot at present be accurately predicted and must be determined empirically by evaluating responses on intact plants. Moreover, in vitro screens are only capable of screening test compounds against a single or very small number of targets, thus requiring separate assays for each new target.
A method for screening compounds for herbicidal or other biological activity which combines the high information content and good predictive qualities of greenhouse screening with the reduced size, cost and compound requirements of in vitro screens, wherein high-throughput screening of test compounds can be accomplished using whole plant responses as the assay, is desirable.
Dornbos and Spencer have examined the germination and early seedling growth of three species of weeds (Medicago sativa, Lolium multiflorum, and Abutilon theophrasti) in 24-well tissue culture plates (Dornbos, D. L. Jr. and Spencer, G. F. J Chem. Ecol. (1990), 16, 339-351). The wells in these plates were cylinders, about 1.6 cm in diameter and 1.7 cm in height, with a total volume of about 3.4 cm.sup.3. Each well contained 1 mL of 0.5% agar in water. Various amounts of test compounds were added to the surface of the agar in 1 mL of solvent (99% hexane, 1% chloroform) and the solvent was then allowed to evaporate. Seeds were placed on the agar and after 3 days, seed germination and seedling length were measured. This assay is terminated at 3 days after inoculation of the seed onto the agar, a point at which the seedlings are still largely dependent on seed reserves for growth. Thus, it is likely that the short term growth will not allow the detection of injury of certain classes of herbicides which affect photosynthetic or other processes which are not required for early seedling growth.